Article pubs.acs.org/IECR
UltraHigh Molecular Weight Nonlinear Polycarbonates Synthesized in Microlayers In Hak Baick,† Yuesheng Ye,† Carla Vanesa Luciani,† Yun Gyong Ahn,‡ Kwang Ho Song,§ and Kyu Yong Choi*,† †
Department of Chemical and Biomolecular Engineering, University of Maryland, College Park, Maryland 20742, United States Seoul Center, Korea Basic Science Institute, 126-16, 5th Street Anam-dong, Seongbuk-Gu, Seoul, Korea 136-701 § Department of Chemical and Biological Engineering, Korea University, Anam-dong, Seongbuk-Gu, Seoul, Korea 136-713 ‡
ABSTRACT: Ultrahigh molecular weight bisphenol A polycarbonate (BPAPC) with nonlinear chain structures has been synthesized by solid-state polymerization in microlayers at 230 °C. It has been found that the molecular weights of the microlayers of low molecular weight prepolymers with thickness ranging from 5 to 35 μm increased rapidly to ultrahigh molecular weight (300,000−600,000 g/mol) without any discoloration. The amorphous prepolymers exhibited much higher reaction rate than the partially crystallized prepolymers. Although the reactive end group mole ratios of the prepolymers prepared by semibatch melt polycondensation deviated significantly from the stoichiometric value, branching and partial cross-linking reactions have been found to occur to rapidly increase the polymer molecular weight. These nonlinear chain structures observed were due to Fries rearrangement, Kolbe-Schmitt rearrangement reactions, and radical-induced scission/cross-linking reactions. The microstructures of the high molecular weight polymers have been analyzed by 13C NMR, 1H NMR, pyrolysis-gas chromatography mass spectrometry, and atomic force microscopy.
1. INTRODUCTION Linear bisphenol A polycarbonate (BPAPC) is an important engineering polymer that is widely used in automotives, electronic displays, data storage, medical, environmental, energy, and aerospace industries.1 The high glass transition temperature (Tg ∼ 150 °C), excellent optical clarity, and high impact resistance are the major advantages of BPAPC.2−5 Molecular weight (MW), molecular weight distribution (MWD), and microchain structures (i.e., linear, branching, and cross-linking) are the key parameters that impact the physical, mechanical, and flow properties of polycarbonate. In particular, the molecular weight of BPAPC is perhaps one of the most important properties in actual applications. For example, relatively low molecular weight BPAPC is suitable as a substrate for digital storage media, whereas high molecular weight BPAPC (35,000−70,000 g/mol) is needed for injection molding applications. BPAPC of molecular weight of about 30,000 g/mol can be used as optical fiber core material for solid-state emitters and detectors with lower infrared absorption.6,7 High transparency in high molecular weight polycarbonate is also an important attribute required for the lamination of BPAPC (∼50,000−100,000 g/mol) with polyvinyl butyral in film applications such as vehicle and building windows, skylights, retail (jewelry) casings, and bulletresistant windows.2,8,9 BPAPC is manufactured industrially by interfacial phosgenation and polycondensation processes (i.e., melt polycondensation and solid-state polymerization). In a melt polycondensation process, bisphenol A (BPA) and diphenyl carbonate (DPC) monomers are reacted reversibly under vacuum above the polymer melting temperature (Tm ∼ 260 °C) in the presence of a catalyst such as lithium hydroxide monohydrate (LiOH·H2O). The byproduct of the reversible condensation © 2013 American Chemical Society
reaction is phenol. To drive the reaction forward to produce linear polycarbonate, phenol must be removed effectively from the high viscosity polymer mixture by applying low pressure (∼0.1 mmHg) or using an inert purge gas. High reaction temperature, long reaction time (4−5 h), and ineffective removal of phenol often lead to unwanted side reactions, causing unwanted discoloration of the final product.10−12 One of the drawbacks of a melt BPAPC process is that obtaining molecular weight higher than 30,000 g/mol is difficult because very high viscosity of polycarbonate melt (e.g., 8,000−20,000 poise at 280 °C) makes the removal of phenol very difficult.13 If the reaction time is extended at high reaction temperature to increase molecular weight, discoloration and gel formation may result.14−17 Although the amount of such reaction products might be very small, their impact on the polymer quality is quite detrimental. To solve these problems that occur in high temperature melt polycondensation processes, a solid-state polymerization (SSP) technique can be used. SSP is a postmelt polycondensation process which is conducted at a temperature lower than the polymer’s melting point but above its glass transition temperature (e.g., 180−235 °C) to obtain high molecular weight polycarbonates with minimum side reactions.18−21 In a typical SSP process, pellets or chips of relatively low molecular weight prepolymer of several hundred micrometers or a few millimeters are partially crystallized and heated to a reaction temperature between the glass transition temperature and the crystalline melting temperature of the polymer to prevent partial melting of polymer particles.13,22,23 Received: Revised: Accepted: Published: 17419
September 6, 2013 November 15, 2013 November 23, 2013 November 23, 2013 dx.doi.org/10.1021/ie4029544 | Ind. Eng. Chem. Res. 2013, 52, 17419−17431
Industrial & Engineering Chemistry Research
Article
in a fume hood for 2 h at ambient temperature and pressure. The coated prepolymer microlayers prepared by this method were transparent and amorphous. The prepolymer microlayers of different thickness (5−35 μm) were obtained by varying the polymer concentration (e.g., 7.0 to 25.0 wt %), and the thickness measurement error was within ±1.3 μm. The microlayer thickness was measured by a Mitutoyo micrometer (Japan). When the microlayer thickness was larger than 35 μm, partial crystallization occurred as the casting solvent evaporated from the sample and the microlayer became opaque. Thus, the maximum thickness of amorphous microlayer samples studied in this work was 35 μm. For the solid-state polymerization, a vacuum oven was used as a reaction chamber (Fisher Scientific Isotemp Model 281A Vacuum Oven, 0.65 cu. Ft.). The polymerization experiments were carried out in a reaction chamber at 230 °C and 10 mmHg. Polymer samples were taken from the reaction chamber at different reaction times. For the analysis, the polymer samples were taken out of the reaction chamber at designated sampling times. 2.2.2. Solid-State Polymerization in Partially Crystallized Microlayers (SSPm). Partially crystallized prepolymer microlayers were prepared by treating the amorphous microlayers with acetone. The resulting partially crystallized polycarbonate microlayers exhibited three-dimensional spherulitic morphology as reported in the literature.26,27 The residual acetone was removed by air and vacuum drying at room temperature for 48 h. The degree of crystallization of the partially crystallized BPAPC microlayers was measured by differential scanning calorimetry (DSC), and it was about 32% for most of the crystallized samples. The polymerization experiments were carried out in a reaction chamber at 230 °C and 10 mmHg. 2.2.3. Solid-State Polymerization of Partially Crystallized Polymer Microparticles (SSPp). To compare the performance of the solid-state polymerization in microlayers (SSPm), we have conducted conventional SSP experiments using the same prepolymers used in the SSPm but in particle form. The prepolymer was first dissolved in chloroform and then precipitated in methanol. The precipitated prepolymer particles of about 10−350 μm-radius were dried in vacuum for 48 h and crystallized in acetone, followed by drying under vacuum at room temperature for 48 h. The degree of crystallinity measured by DSC was about 33%, which was quite similar to that of the partially crystallized microlayers (32%). The scanning electron microscopy (SEM) analysis of these crystallized BPAPC particles showed that they were the aggregates of smaller precipitated particles. The particles were classified using sieve trays. The conventional SSP was performed using the same reaction conditions as in SSPm and the reaction chamber used in the microlayer polymerization. Details of the experimental procedure for the conventional SSP can be found elsewhere.20−23,28 2.3. Polymer Characterization. The polymer molecular weight and molecular weight distribution (MWD) were measured using a gel permeation chromatography (GPC) system (Waters) equipped with PLgel 10 μm MIXED-B columns (Polymer Laboratories) and a UV detector (Waters 2487). Chloroform was used as a mobile phase. Polystyrene standards of narrow molecular weight distributions were used to construct a molecular weight calibration curve. The end group mole ratio in a prepolymer sample (i.e., phenyl carbonate end group ([−OCO−C6H5])/phenolic end group ([−OH])) was determined by 13C NMR spectroscopy (Bruker DRX-500 spectrometer operating at 500 MHz using deuterated chloro-
The polymer molecular weight can be increased, but the reaction is very slow and long reaction time is required. In this work, we report new experimental studies on the polymerization of BPAPC in thin microlayers (5−35 μm in thickness) at the similar temperature range of solid-state polymerization of BPAPC. The dimension of the microlayers employed in this study is far smaller than the dimension of polymer commonly used in typical solid-state polymerization. This paper illustrates that the polymerization of BPAPC in amorphous microlayers shows quite unusual polymerization behavior which does not follow the kinetic mechanism of a classical linear polycondensation process for the synthesis of polycarbonate.
2. EXPERIMENTAL SECTION 2.1. Materials. The chemical reagents including bisphenol A (BPA, 99.9%, Aldrich) and diphenyl carbonate (DPC, 99%, Aldrich) were recrystallized using methanol and water solution (1:1 v/v) and methanol, respectively. Lithium hydroxide monohydrate (LiOH·H2O, Aldrich) was used as received. The low molecular weight PC prepolymers were prepared by semibatch melt polycondensation using BPA and DPC as monomers and LiOH·H2O as a catalyst (1.75 × 10−4 M [LiOH·H2O]/[BPA]) following the polymerization procedure reported in the literature.24,25 The polycarbonate prepolymers of three different molecular weights and the end group mole ratios (ra′) (i.e., phenyl end group ([−OCO−C6H5])/phenolic end group ([−OH])) were used, and their properties are shown in Table 1. The prepolymers were used without removing the catalyst. Table 1. Properties of BPAPC Prepolymers prepolymer
M̅ w (g/mol)
M̅ w/M̅ n
ra′ =[−OCO−C6H5]/[−OH] (mol/mol)a
A-8k B-14k C-21k
8,300 14,000 21,000
1.95 1.97 2.00
0.66 0.80 0.90
a
Measured by 13C NMR.
2.2. Preparation of Amorphous, Crystalline Microlayers, and Crystalline Microparticles. The following polymerization experiments were carried out: (i) polymerization in amorphous polymer microlayers; (ii) polymerization in partially crystallized polymer microlayers; and (iii) conventional solid-state polymerization (SSP) of partially crystallized prepolymer particles for the purpose of comparison. Since the catalyst was not removed from the prepolymers, the catalyst content in each sample for the microlayer polymerization was the same as the prepolymer produced from the melt polycondensation. Details of the polymerization procedures are described as follows. 2.2.1. Solid-State Polymerization in Amorphous Microlayers (SSPm). Amorphous microlayer polymer samples were prepared using a solvent casting technique. First, a predetermined amount of low molecular weight prepolymer sample (Table 1) was dissolved in a solvent (chloroform) at room temperature. Silica substrate (2.5 cm × 7.5 cm) was cleaned with acetone for 10 min and then cleaned again with methanol for 5 min before being rinsed with deionized water and dried by nitrogen gas blow. The cleaned silica substrate was preheated and immersed in a bath of prepolymer solution and removed. The silica substrate coated with prepolymer was dried 17420
dx.doi.org/10.1021/ie4029544 | Ind. Eng. Chem. Res. 2013, 52, 17419−17431
Industrial & Engineering Chemistry Research
Article
form (CDCl3) as a solvent). We analyzed the polycarbonate molecular architecture using high resolution 13C NMR and 1H NMR spectroscopy (Bruker AV III spectrometer at 150.9 and 600 MHz, respectively) and pyrolysis-gas chromatography mass spectrometry (Py-GC/MS). For the 13C NMR and 1H NMR spectroscopy analysis, polymer samples taken from the SSPm experiments were dissolved in deuterated chloroform (CDCl3). For the Py-GC/MS analysis, we used a Frontier Lab PY-2020iD pyrolyzer (Koriyama, Fukushima, Japan), connected to an Agilent GC/MS system composed of an Agilent 6890 gas chromatograph and an Agilent 5975i mass spectrometer (Palo Alto, CA, USA), to separate and obtain the mass spectra of the compounds in each sample. Measurement conditions for the GC were as follows: A DB-5 MS capillary column (length, 30 m; internal diameter, 250 μm; film thickness, 0.25 μm, 5% diphenyl-95% dimethylsiloxane phase, J&W Scientific, Folsom, CA, USA); carrier gas helium running at a constant flow rate of 1 mL/min (37 cm/s); split mode (50:1 ratio). The initial temperature of the oven was 60 °C, and then a 10 °C/min gradient was applied to 320 °C (15 min). The column was interfaced directly to the electron impact (EI) ion source of the mass spectrometer. The ion source was operated at 70 eV. The injection port, transfer line, and ion source temperature were set at 300, 300, and 230 °C, respectively. The mass spectrometer was scanned in the 50−800 m/z range. An aqueous solution (25 wt %) of tetramethylammonium hydroxide [(CH3)4NOH; TMAH] (Aldrich) was used as the reagent for methylation of the product samples. About 100 μg of the sample and 10 μL of the aqueous solution of TMAH taken in a platinum sample cup were introduced into the furnace at 600 °C. Then, the temperature program of the gas chromatograph oven was started. Atomic Force Microscope (AFM D-3000) with a high resolution diamond-like tip (Hi′RES-C14, MicroMasch) was also used to obtain the images of single polycarbonate molecules. The thermal characteristics of BPAPC were analyzed by differential scanning calorimetry (DSC) (TA Instruments Q1000) with a heating rate of 10 °C/min.
Figure 1. Weight-average molecular weight vs reaction time profiles (T = 230 °C, P = 10 mmHg, prepolymer B-14k): (a) (⧫) SSPm (10 μm layer, amorphous); (▲) SSPm (10 μm layer, crystalline). Solid and dashed lines represent the numerical simulations for the crystallized microlayer (10 μm-thick) and amorphous microlayer (10 μm-thick), respectively; (b) (●) SSPp (10 μm radius particles, crystalline); (■) SSPp (350 μm radius particles, crystalline). Solid lines represent the numerical simulations.
polymer microlayer was partially crystallized (▲), the polymer molecular weight increased to a much lower value (about 125,000 g/mol) in 180 min, although the molecular weight of 125,000 g/mol is a very high value. Figure 1 (b) shows the molecular weight profiles when the solid-state polymerization was carried out using polymer microparticles. Note that when the prepolymer microparticles of radius 10 μm (●) were used with the same reaction conditions, the molecular weight increased to about 96,000 g/mol, which is slightly lower than the case of the solid-state polymerization in partially crystallized microlayers of 10 μm thickness (▲ in Figure 1(a)). However, when larger crystallized polymer particles (r = 350 μm) (■) were used, molecular weight increased only to 26,500 g/mol in 180 min. The size of this polymer particle represents a typical particle size employed in conventional solid-state polymerization processes. We have also observed that the final polymer microlayers were highly transparent without any discoloration. Discoloration of polycarbonate is known to be one of the problems in high temperature melt transesterification processes.11,12 For example, in conventional melt polycondensation processes at 260−280 °C and low pressure, the employment of long reaction time to obtain high molecular weight (30,000−60,000 g/mol) often leads to unwanted discoloration due to some unwanted side reactions with phenol (e.g., Kolbe-Schmitt type) that remains in the viscous reaction mixture if not completely removed by vacuum. In order to develop some insight into the unusual characteristics of the SSPm in amorphous polymer microlayers,
3. RESULTS AND DISCUSSION 3.1. Solid-State Polymerization in Microlayers and Microparticles. The first series of experiments were carried out using the prepolymer sample B-14K (Table 1) at 230 °C and 10 mmHg for 180 min in four different settings: (i) SSPm in 10 μm-thick amorphous microlayers, (ii) SSPm in 10 μmthick partially crystallized microlayers, (iii) SSPp (solid-state polymerization in particles) in partially crystallized particles of 10 μm-radius, and (iv) SSPp in partially crystallized particles of 350 μm-radius (typical particle size in SSP). Here, the microlayer thickness (10 μm) is a nominal value, and the actual layer thickness varies within ±1.3 μm. Figure 1 shows the experimentally measured weight-average molecular weight (Mw) vs reaction time profiles for these four representative samples (symbols). The most prominent result shown in Figure 1(a) is that the molecular weight of 10 μmthick amorphous prepolymer microlayers (⧫) increased from 14,000 g/mol (prepolymer MW) to 340,000 g/mol in 180 min. Although a postmelt polymerization process such as solid-state polymerization can be used to raise the molecular weight of BPAPC to 36,000−42,000 g/mol in 16−24 h at the temperature of 230 °C,19,21 the rapid increase to such a high molecular weight as illustrated in Figure 1 for BPAPC has never been reported in the literature. It is also seen that when the 17421
dx.doi.org/10.1021/ie4029544 | Ind. Eng. Chem. Res. 2013, 52, 17419−17431
Industrial & Engineering Chemistry Research
Article
molecular weight. For an AA-BB type of polycondensation process (e.g., bisphenol A polycarbonate synthesis from bisphenol A and diphenyl carbonate), any deviation from the stoichiometric end group ratio (ra′ = [−OCO−C6H5]/[−OH] = 1.0) is unfavorable for the growth of polymer chain length. For example, when BPAPC prepolymer is synthesized from BPA and DPC at high reaction temperature and reduced pressure, a partial loss of volatile DPC occurs, making the end group ratio difficult to maintain at its stoichiometric value in the reaction mixture during the course of polymerization.23,29 To compensate for the loss of DPC during the polymerization, a slight excess amount of DPC is generally used. Indeed, the three prepolymer samples we prepared show the end group ratios deviating from 1.0 (Table 1). We used these three prepolymer samples to investigate the effect of end group ratio on the rate of polymerization in microlayer. Figure 3 shows the increase of molecular weights as a function of reaction time. Here, three amorphous prepolymer
we have carried out the numerical simulations of the solid-state polymerization using the solid-state polymerization model reported in refs 29−31. The model equations are provided in the Appendix. The kinetic model that consists of the population balance equations for the molecular species defined by the end group type is incorporated into a dynamic diffusion and reaction model of a particle. In the polymer particle model, the degree of crystallinity has also been considered. The original solid-state model developed for a spherical model has been modified for the simulation of a microlayer system. Model simulation results for the end-group ratios of 0.80 (i.e., prepolymer sample B) are shown in Figure 1. The modeldata match is quite good for the crystallized particles (Figure 1(b), solid lines) and crystallized microlayers (Figure 1(a), solid line), indicating that the model and the model parameters used in the simulation were quite satisfactory for the partially crystallized polymer layers or particles. However, the model predictions of molecular weight for the amorphous microlayers (dashed lines in Figure 1(a)) show significant deviations from the experimental data after 30 min of polymerization. It suggests that the solid-state polymerization in amorphous microlayers may not be adequately described by the classical diffusion and reaction model with a linear step-growth polymerization mechanism. The molecular weight distributions (MWDs) of prepolymer and three polymerized samples are shown in Figure 2. We
Figure 3. Effect of prepolymer end-group ratio on the evolution of weight-average MW in amorphous polymer microlayers of 10 μm thickness (T = 230 °C, P = 10 mmHg): amorphous microlayer (solid line) and crystallized microlayer (dashed line). Lines were added to guide the eyes only.
Figure 2. Molecular weight distributions (MWD) of microlayer and particle samples (T = 230 °C, P = 10 mmHg, prepolymer B-14k) at 180 min: a, 10 μm amorphous microlayer; b, 10 μm crystalline microlayer (acetone induced crystallization); c, 10 μm-radius crystalline particles.
microlayers and three partially crystallized prepolymer microlayer samples were used, and the thickness of each microlayer was kept constant at 10 μm within experimental error. As observed in the prepolymer sample B-14K (Figure 1), all other samples show the similar effects of rapid molecular weight increase even with the stoichiometric imbalance. The amorphous sample C-21k, which has the end group mole ratio ([−OCO−C6H5]/[−OH]) closest to 1.0 among three samples tested (Table 1), gives the highest molecular weight of all as expected. Notice that, after 150 min of reaction, the amorphous microlayer of prepolymer sample C-21k has reached a molecular weight of 600,000 g/mol. Such a high molecular weight and soluble polycarbonate has not been reported in the literature. Even the crystallized microlayers of sample C-21K have the molecular weight about 300,000 g/mol after 240 min. 3.3. Effect of Amorphous Microlayer Thickness. The thickness of a microlayer is expected to impact the molecular weight increase phenomena because it affects the diffusion rate of phenol (condensation byproduct) from the polymer microlayer. Figure 4 shows the molecular weight vs reaction time profiles for the amorphous microlayers of thickness ranging from 5 to 35 μm. All these samples were prepared by solvent casting, and they maintained transparency during the
observe that the amorphous microlayers (data marked by ⧫ in Figure 1a) that have the highest molecular weight of all (peak a) show the presence of polymer chains having molecular weight larger than 1 million g/mol (∼10 wt %). Also, we observed a small increase in the polydispersity (PD = M̅ w/M̅ n) from 1.97 (prepolymer B-14k) to 2.17 (10 μm-radius crystalline particles), 2.15 (10 μm crystalline layer), and 2.57 (10 μm amorphous layer). According to the linear step-growth polymerization theory, the polydispersity is close to 2.0 even for very high molecular weight polymers. Thus, the MWD broadening (i.e., deviations from PD = 2.0) observed in our experiments suggests that some deviation from the homogeneous linear step-growth polymerization might have occurred. For example, a spatial distribution of phenol due to diffusion resistance might have been present or the polymer chain structures might have deviated from the linear configuration. We shall discuss this issue later in this paper. 3.2. Effect of End-Group Ratio. In a linear step-growth polymerization, the end-group ratio is an important parameter that affects the rate of polymerization and the polymer 17422
dx.doi.org/10.1021/ie4029544 | Ind. Eng. Chem. Res. 2013, 52, 17419−17431
Industrial & Engineering Chemistry Research
Article
ratio ([−OCO−C6H5]/[−OH]) in the starting prepolymer. Also, we note that the polymer MWD broadening has also occurred with the progress of reaction. To understand the unusually high molecular weight in our polymer samples, particularly in amorphous microlayers, we postulate that the polymer chain structure may not be perfectly linear but may have some nonlinear structures such as branching and partial cross-linking. In fact, in a melt polymerization of polycarbonate at high reaction temperatures (260−290 °C), a small amount of branched polycarbonate can be formed by Kolbe-Schmitt rearrangement or Fries rearrangement reactions.32,33 Although the concentration of branched polymers (often called Fries product) in linear polycarbonates is usually quite small (∼450 ppm), they can adversely affect melt flow properties and ductility, and hence many efforts have been made to minimize the formation of Fries product.32−36 The large polydispersities observed in our polymers from amorphous microlayer (M̅ w/M̅ n = 2.57) indicate that the presence of branched polymers might be a strong possibility. Thus, we have analyzed the polycarbonate molecular architecture to probe the presence of branched polymers using high resolution 13C NMR and 1H NMR spectroscopy (Bruker AV III spectrometer at 150.9 and 600 MHz, respectively). The assignments of carbon and proton peaks were based on the literature data.32,37−39 Scheme 1 illustrates the reaction mechanisms that lead to the branched structures A and B through Kolbe-Schmitt rearrangement or Fries rearrangement reactions. The reaction mechanisms leading to the formation of cross-linked chain structures are shown in Scheme 2. Figure 6 shows the 13C NMR spectral data for an amorphous prepolymer (B-14k) and three amorphous samples, all in 10 μm microlayers, taken at different reaction times (15 min, 30 min, and 60 min). Here, we note that the phenolic group and the phenyl group decreased to insignificantly low levels after 30 min. However, according to the polymerization data in Figure 1 (⧫), polymer molecular weight continued to increase after 30 min, suggesting that some other reactions might have led to the continued growth of polymer chains and molecular weight. The 1H NMR spectroscopy data of prepolymer, high molecular weight amorphous microlayer samples, and crystalline microlayer sample are shown in Figure 7. For the amorphous polymers, the aromatic regions (peak assignment 3‴ at 8.0−8.1 ppm) for phenyl salicylate phenyl carbonate (PhSALPhC, structure B in Scheme 1) and peak assignment 6″ at 2.1−2.2 ppm (structure C in Scheme 2) indicate the presence of anomalous chain structures (branched and crosslinked structures) in the high molecular weight samples obtained after 30 min. Note that these peak assignments were absent in the low molecular weight prepolymers and crystalline microlayers used in our study (see Figure 7). In the literature, the presence of the branched units has been reported.32 For example, the concentration of PhSALPhC structure units (i.e., branching density) in the fractionated polycarbonate samples of molecular weight (Mw) of 11,400− 39,200 (g/mol) was reported to increase quasi linearly with molecular weight, suggesting that the branching units were not homogeneously distributed across the molecular weight distribution. Although the sample molecular weights in this reference are much lower than those obtained in our experiments, we performed the following simple calculations: The literature data32 can be correlated to Y = 0.1125Mw + 822.5 where Y is the modified concentration of PhSALPhC (in ppm) and Mw is
Figure 4. Effect of the amorphous microlayer thickness on the evolution of the polymer molecular weight with the reaction time (T = 230 °C, P = 10 mmHg, prepolymer B-14k). Lines were added to guide the eyes only.
entire period of polymerization. When the microlayer thickness was larger than 35 μm, partial crystallization occurred during the solvent casting, and, hence, we limited the microlayer thickness below 35 μm to ensure that prepolymer microlayers were amorphous. As expected, Figure 4 shows that within the range of experimental conditions, polymer microlayer thickness has a strong effect on the molecular weight. For the microlayer thickness ranging from 5 to 20 μm, the polymer molecular weight increases almost linearly as the layer thickness is reduced; but 35 μm microlayer samples show slow increase in molecular weight to the final value of 60,000 g/mol in 180 min. We have also observed that for the microlayer thickness range of 5−20 μm, a small amount of insoluble fraction of polymer was obtained after 180 min (e.g., 4.8 wt % in 10 μm thickness sample), while no insoluble PC was obtained for the 35 μm samples that exhibited lower molecular weight. Figure 5 shows the UV−vis spectra of prepolymer (B-14k) and amorphous microlayer polymer samples taken at different
Figure 5. UV−vis spectra of (a) prepolymer (B-14k) and (b-e) amorphous microlayers with reaction time (T = 230 °C, P = 10 mmHg, prepolymer B-14k).
reaction times. In the wavelength range of 300−800 nm, 89− 94% of transmission has been observed. 3.4. Characterization of the Polymer Structures and Properties. 3.4.1. 13C NMR and 1H NMR Spectroscopic Analysis. In the foregoing discussion, we have seen that the polymer molecular weight in microlayers increased rapidly even with a significant stoichiometric imbalance of the end group 17423
dx.doi.org/10.1021/ie4029544 | Ind. Eng. Chem. Res. 2013, 52, 17419−17431
Industrial & Engineering Chemistry Research
Article
Scheme 1. Kolbe-Schmitt and Fries Rearrangement Reactions Leading to Branched Structures in Polycarbonates32,33
Scheme 2. Reaction Mechanisms for the Formation of Cross-Linked Polycarbonatesa
a
• represents a macroradical species generated in the system.33
the weight average molecular weight (in g/mol). According to this correlation, as polycarbonate molecular weight increases
from 14,000 (prepolymer B-14K in our experiment) to 40,000 g/mol (t = 30 min), the branching unit concentration increases 17424
dx.doi.org/10.1021/ie4029544 | Ind. Eng. Chem. Res. 2013, 52, 17419−17431
Industrial & Engineering Chemistry Research
Article
from 0.24% to 0.5%. If the correlation is also applied to a higher MW sample (e.g., 340,000 g/mol) as observed in our study (e.g., Figure 1), the calculated branching unit concentration is as large as 3.8%, which is about 16 times the unit concentration in the prepolymer. It is to be noted that none of the peaks in Figures 6 and 7 shows the presence of phenol, indicating that phenol was effectively removed from the polymer microlayers. Another possible nonlinear chain structure in the high molecular weight BPAPC obtained in our study is a partial cross-linking. The formation of a cross-linked structure in polycarbonate during melt polymerization has been known to be possible, and it was proposed that the cross-linking can occur by radical recombination as illustrated in Scheme 2.33 The reaction scheme includes the hydrogen abstraction and scission of oxygen and carbonyl linkage to generate radicals that lead to the cross-linked structures through recombination.33,38 The mole fractions of cross-linkages (peak assignment 6″ at 2.1−2.2 ppm) in sample B-14k from amorphous microlayers at 230 °C, estimated using the area intensity of methyl proton and ethyl proton in repeating unit and cross-linkage, were 3.55 × 10−4 and 4.26 × 10−2 at 30 min and 180 min, respectively, indicating that the cross-linkages have increased about 2 orders of magnitude in 150 min of solid-state polymerization in microlayers. Table 2 shows the mole ratio of methyl protons to aromatic protons in the amorphous and crystalline polycarbonate microlayer samples at different reaction times. It is seen that while the ratio remains constant in the 10 μm thick crystalline microlayers, the ratio decreases with the progress of reaction when the amorphous microlayers of 10 μm thickness were polymerized. If only linear step growth polymerization occurs,
Figure 6. 13C NMR spectroscopic analysis of the amorphous microlayer samples at different reaction times (T = 230 °C, layer thickness = 10 μm, prepolymer B-14k).
Figure 7. 1H NMR spectroscopic analysis of the amorphous microlayer (T = 230 °C, layer thickness = 10 μm, prepolymer B-14k). Arom. indicates aromatic protons. 17425
dx.doi.org/10.1021/ie4029544 | Ind. Eng. Chem. Res. 2013, 52, 17419−17431
Industrial & Engineering Chemistry Research
Article
Table 2. Ratio of Methyl Protons to Aromatic Protons in Polycarbonates in Partially Crystallized Microlayers (10 μmCrystalline) and Microlayers (10 μm-Amorphous) at 230 °C (P = 10 mmHg, Prepolymer B-14k) ratio of methyl protons to aromatic protonsa 10 μm-crystalline
10 μm-amorphous
time (min)
microlayer
Mw (g/mol)
microlayer
Mw (g/mol)
0 30 60 180
0.716 0.713 0.714 0.716
14,250 34,537 59,470 123,061
0.716 0.703 0.639 0.598
14,250 38,535 109,427 338,347
a
Ratio of methyl protons to aromatic protons = (sum of methyl protons 6, 6′ peak area)/(sum of aromatic protons 2, 2′, 2″ and 3, 3′, 3″ peak area). Figure 9. Mass spectrum of peak 1. Peak 1 generated via reactive pyrolysis from structures A and B at 600 °C.
this ratio will remain constant, but if radical recombination reactions occur (see structure C, D, and E in Scheme 2), this ratio will decrease. The assignment of proton peaks used in Table 2 is provided in Figure 7. 3.4.2. Pyrolysis-Gas Chromatography Mass Spectrometry (Py-GC/MS). The formation of anomalous chain structures such as branched and cross-linked structures in the amorphous microlayer samples has been investigated using pyrolysis gas chromatography mass spectrometry (Py-GC/MS). Py-GC/MS technique enables the characterization of macromolecular complexes, and it has been found to be a very effective technique for the qualitative analysis of various condensation polymers. In this method, the polycarbonate chains in the presence of tetramethylammonium hydroxide (TMAH) decompose selectively at high temperatures (e.g., 400 °C) at carbonate linkages to yield methyl derivatives of the components for a given polymer sample. The mass spectrum of each peak provides the identity of a specific chain structure of the polycarbonate sample. Figure 8 shows the pyrograms obtained by the reactive pyrolysis of the prepolymer and the amorphous microlayer
Figure 10. Mass spectrum of peak 2. Peak 2 generated via reactive pyrolysis from structure E at 600 °C.
Figure 8. Pyrograms of prepolymer (B-14k) and amorphous microlayer sample taken at 30 min in the presence of TMAH.
samples in the presence of TMAH at 600 °C. Here, we note four peaks (1, 2, 3, 4) that correspond to anomalous polycarbonate chain structures.33 Mass spectra of these four peaks are shown in Figures 9, 10, 11, and 12. These peaks 1−4 are more pronounced in the sample obtained by polymerization of amorphous microlayers than in the prepolymer. In what follows, we shall present the analysis of these peaks and relevant reaction pathways and resulting chain structures. Rearrangement of the carbonate group in BPAPC can form a pendant carboxyl group, ortho to an ether link, and through
Figure 11. Mass spectrum of peak 3. Peak 3 generated via reactive pyrolysis from structure C at 600 °C.
ester exchange reaction with another BPAPC chain, it can form a branching structure (the Kolbe-Schmitt rearrangement, Scheme 1(a)). The branching structure can also be formed by the Fries rearrangement reaction (see Scheme 1 (b)). The possibility of homolytic scission of the carbonate group at high temperature between the oxygen and carbonyl group in BPAPC 17426
dx.doi.org/10.1021/ie4029544 | Ind. Eng. Chem. Res. 2013, 52, 17419−17431
Industrial & Engineering Chemistry Research
Article
isopropylidene group and the aromatic ring of structure D at high reaction temperature. Thus, peak 2 detected by Py-GC/ MS (Figure 8) is a supporting evidence of structure D formed by the recombination of methylene radical and phenoxy radical. The formation of D is also confirmed by the mass spectrum of peak 4 shown in Figure 12. To assess the efficiency of phenol removal from the microlayers, we have used a numerical simulation model (Appendix) for the amorphous and partially crystalline polymer microlayers, and the simulation results are illustrated in Figure 13. First of all, it is seen that for a given initial phenol
Figure 12. Mass spectrum of peak 4. Peak 4 generated via reactive pyrolysis from structure D at 600 °C.
and hydrogen abstraction in the system can generate radicals that may lead to a cross-linking reaction (see Scheme 2).33 The branching and cross-linking studies on polycarbonate reported in the literature were mostly on the thermal degradation of polycarbonate in the temperature range 300−500 °C, which is far higher than the polymerization in microlayers employed in our study (
